Open-source AI-powered learning operations. Runs in your CLI. Your data, your machine.
"Companies use AI to filter candidates. EduOrchestrate gives learners AI to build a daily, sourced path to the job."
eo Built for learners who want proof, not passive curriculum.
What is EduOrchestrate?
EduOrchestrate is an open-source, AI-powered learning system that runs locally inside your coding agent. It captures your profile, target role, and first-skill focus, then builds an adaptive plan with a fixed 30-day minimum, installs agent adapters, gathers role-specific course sources, runs a token-conscious harness, and turns every day into one build artifact with visible proof: research current references, practice one skill, preview or send the daily email, and log evidence. Your data, your machine — no cloud, no dashboard, no subscription. Scheduled delivery runs through GitHub Actions and stops after the plan window. Open source, MIT-licensed, and free forever.
Try it out
npx eduorchestrate
$ npx eduorchestrate
$ npx eduorchestrate plan --role "Agentic AI and LLM Engineer" --skill "RAG evaluation"
$ npx eduorchestrate research --day 1
$ npx eduorchestrate send-today --dry-run
$ npx eduorchestrate setup-secrets
$ npx eduorchestrate send-todayInstead of collecting courses in a spreadsheet, you get an AI-powered daily loop that plans, researches, emails, and tracks visible learning proof.
"It's like having a mentor for your job search skills, but without the subscription dashboard."
-- EduOrchestrate daily loopBuilt for the daily loop.
Plan, research, ship, prove. Each card below is a command you can run today.
00integrations
Works with any coding agent that can read project instructions and MCP-capable adapters. One repo, every assistant.
Claude Code
Codex
Gemini
OpenClaw
GitHub Actions
MCP
01build
Every plan day includes one practical task, one proof target, review questions, and a next-day bridge. You ship evidence instead of reading endless curriculum lists.
See role blueprints ->02deliver
Dry-run previews stay manual. Real scheduled delivery uses GitHub Actions secrets, computes the current day from schedule.startDate, and skips after the plan is complete.
03prove
Run progress-card or terminal-card to render your role, first skill, active day, next action, latest evidence, and momentum.
State stays local in data/. SMTP secrets stay in .env.local or GitHub Actions secrets. Nothing phones home.
What can it plan?
Pick a target role; the planner expands it into a sourced, 30-day-minimum build path.
Hugging Face, GitHub, Colab, RAG, evaluation, agents, and portfolio-ready proof.
Frontend, backend, APIs, auth, testing, deployment, and a shippable dashboard.
SOC thinking, SIEM, logs, MITRE ATT&CK, incident notes, and detection projects.
Python notebooks, pandas, SQL, modeling, evaluation, and stakeholder summaries.
Any other role works too — these are just the tuned blueprints.
EduOrchestrate is not a course marketplace or a cloud LMS. It is a local command loop for learners who want daily proof.
It starts from your target role, current learning topic, first skill focus, current stage, weekly hours, and requested plan length. Plans never go below 30 days.
Only after one-time setup. Manual agent runs should preview first. Scheduled GitHub Actions runs can send automatically using stored secrets and the configured schedule.
No. The scheduled path is a GitHub Actions workflow or another runner. It does not rely on a browser instance or an agent session continuing in the background.
Plans, progress logs, research digests, and cards are written to data/. SMTP credentials live in .env.local or GitHub Actions secrets.
Claude Code, Codex, Gemini, OpenClaw, Antigravity, Hermas, generic coding agents through AGENTS.md, and MCP-capable agents.
$ npx eduorchestrate
Or follow what ships next.
Release announcements and occasional updates. Unsubscribe anytime.